Wall Observer BTC/USD - Bitcoin price movement tracking & discussion

[oda.krell]

Alright... Let me be clear that I'm not gloating. I really wouldn't mind seeing a reversal either finally.

That said, the situation is starting to look pretty different from July 18 (which I guess some of us were hoping for)... Back then: bottom, sharp rally, sharp drop, found support, next rally. This time, it's more like: Bottom, quick rally, sharp drop, little support, fizzle out.

We need fresh fiat. Maybe I'll start walking around with a sandwich board in town that says 'Invest in Bitcoin now' :/

[p0peji]

Maybe I should remove my 425 bids

Should I?

For a quick buck maybe, but I would not hold on to them for too long.

[wachtwoord]

What's the reason for the dumps?

China ban?

Which ban? They are banning bitcoin every 2-3 weeks since December, get over it.

The butthurt is strong in this one  

Indeed: what ban?

[wachtwoord]

Maybe I should remove my 425 bids

Should I?

For a quick buck maybe, but I would not hold on to them for too long.

For a quick buck I don't know, but for long term hold it's perfect

[rpietila]

Though the great thing about linear regression is that as long as the price starts out at 0, the slope can't ever be negative, so all is well, and all will forever be well so long as there is a non-zero trading price. Infinite profits ahead, guaranteed.

Eg.with Auroracoin the slope is already negative. Linear regression could not prevent it.

Hm, that points out a problem with using linear regression on all the monthly points.

Suppose two items A, B that start out both at the same price 1$. Item A then grows steadily to 1000$ over 2 years.  Item B shoots up to 2000$ on the first month and then decays steadily to 1000$ in the next 23 months.

Linear regression will give a positive slope for item A and a negative slope for item B, even though both increased by a factor of 1000 over that time span.

On the other hand, consider item C that starts out at 1000$, drops to 1$ on the first month and recovers steadily to 1000$ over the next 23 months.  Linear regression would give positive slope.

These examples do not prove that linear regression is "wrong", but should make one aware that its meaning is not what one may think.

"Your results are not right. They are not even wrong." --some famous scientist reviewing the work of some colleague

Do you happen to like it in the academia when somebody who has no experience on the subject that you have been researching for months, uses childish non-real world examples to prove that in some cases the model you have been researching and using with incredible precision, would give strange results?

I consider such a person a menace, but luckily Bitcoin will teach him.

[p0peji]

Maybe I should remove my 425 bids

Should I?

For a quick buck maybe, but I would not hold on to them for too long.

For a quick buck I don't know, but for long term hold it's perfect

Long term, you mean for a few years?

[N12]

Bitcoin will teach him.

The cult has really come far.

[Mythul]

2014 will be remember as Bitcoin China drama year !

[wachtwoord]

Maybe I should remove my 425 bids

Should I?

For a quick buck maybe, but I would not hold on to them for too long.

For a quick buck I don't know, but for long term hold it's perfect

Long term, you mean for a few years?

Long term is at least a few years yes.

[oda.krell]

It has already been shown the the model is useful, and that it is in fact reasonable to infer that bitcoin's value grows as the square of its generalized user base.  Realizing that I am using the terms "bitcoins value" and "generalized user base" in a technical way, the task moving forward in time is to find ways to estimate N that simultaneously support the Metcalfe model without stretching the common-sense meanings of the words "value" or "user base."  If we have to stretch common-sense too far, then the model will no longer be useful.  

That's the relevant point here, though. The "usefulness" i.e. the explanatory power, is your motivation for using no. of addresses for N. So far that worked out pretty well. There's not that much history however to test that usefulness on so far, I hope you will agree. If in 3 years it'll turn the no. of addresses was a decent estimate for the initial phase of the market, but now it is not anymore, because it inflates the estimate (or underestimates it), you wouldn't be too surprised I guess. But I see your point, also: you are not commited to this particular way of estimating N, and are willing to replace it with whatever gives the best results (while still being motivated enough to count as a proxy for user base)

[m3g4tr0n]

$15 drop

que the "OMG whats the news!? china again?" posts

[KeyserSoze]

$15 drop
que the "OMG whats the news!? china again?" posts

The catalyst wasn't China this time. This is the market beginning to react to chessnut's "$4000 before $400" bet.

[bitcoinsrus]

$15 drop
que the "OMG whats the news!? china again?" posts

The catalyst wasn't China this time. This is the market beginning to react to chessnut's "$4000 before $400" bet.

lol

[Krabby]

I think loaded took him up on that bet.

[Peter R]

It has already been shown the the model is useful, and that it is in fact reasonable to infer that bitcoin's value grows as the square of its generalized user base.  Realizing that I am using the terms "bitcoins value" and "generalized user base" in a technical way, the task moving forward in time is to find ways to estimate N that simultaneously support the Metcalfe model without stretching the common-sense meanings of the words "value" or "user base."  If we have to stretch common-sense too far, then the model will no longer be useful.  

That's the relevant point here, though. The "usefulness" i.e. the explanatory power, is your motivation for using no. of addresses for N. So far that worked out pretty well. There's not that much history however to test that usefulness on so far, I hope you will agree. If in 3 years it'll turn the no. of addresses was a decent estimate for the initial phase of the market, but now it is not anymore, because it inflates the estimate (or underestimates it), you wouldn't be too surprised I guess. But I see your point, also: you are not commited to this particular way of estimating N, and are willing to replace it with whatever gives the best results (while still being motivated enough to count as a proxy for user base)

Yes, exactly.  

The two ways I am presently estimating N were suggested by Aminorex and someone else when we made this discovery in Risto's TA thread.  Moving forward, we can use anything for N but the more we have to stretch reality to say it bears any relation to what we mean when we say "user base," then the less useful the model becomes.  

If bitcoin adoption (especially as a medium of exchange) continues to grow, the model must breakdown at some point for the proxy N = "number of transactions per day excluding popular address."  For example, if I plug Visa transaction quantities into my model I get bitcoin valuations in the hundreds of million or billions lol.  But this doesn't mean that the model is necessarily broken, just that the old proxy is no longer valid.      

But back to the debate about Risto's regression, I do think it is very interesting how various proxies for N seem to deviate less from exponential growth than bitcoin's market cap.  So I think it is fair to say that the exponential growth model has held so far.  How far into the future it will continue to hold is speculation (as I know you agree).  

[ChartBuddy]

Explanation

[xulescu]

Bitcoinwisdom is completely fubar on Huobi.

Do they have a 90-minute engine "optimization"?
Why is the orderbook shifted by 60 CNY?
Sawtooth price soon?
Free replays on "greatest hits"?
Should we trade with confidence already?
De-escalate?

[prophetx]

well worst case scenario if the price keeps falling pizza.fr accepts payments in bitcoin...

[ChrisML]

Fucking China and all that BS.

When in the fucking fuck fuck can we just move on. All this bullshit with news and dumps. Idiots first class, first class morons. I'd expect this kind of behaviour with alt-coins, not BTC.

I am about to stop checking out bitstamp or BTC forums for the whole month... and return to see what has changed, cause this shit just makes me blood boil.

[bitcoinsrus]

well worst case scenario if the price keeps falling pizza.fr accepts payments in bitcoin...